Skip to content
Branch: master
Commits on Apr 21, 2016
  1. Add meta header to HTML docs to force utf-8

    bapike committed Apr 21, 2016
    When using Firefox 45.0.2 on Debian linux, I noticed that
    Gröbner appears correctly on
    but garbled on
    Changing Firefox's "Text encoding" setting from "Western" to
    "Unicode" fixes the matter, so it appears to be using the
    wrong character encoding.
    All of M2's HTML documentation declares itself as XHTML and
    tries to declare its character encoding as utf-8 in the
    <?xml...> tag.  This apparently works if the XHTML is parsed
    as XHTML.  However, many XHTML documents are not parsed as XHTML;
    for instance, as of April 2016, the web server for M2's
    online documentation includes the HTTP header
      Content-Type: text/html; charset=utf-8
    and this MIME type apparently forces many browsers to read
    the docs as HTML:
    For HTML, it is recommended:
    to always set the character encoding using a <meta...> tag
    or an HTTP header.  Because (I think) of the above problems
    with XHTML being interpreted as HTML, the same is recommended
    for XHTML documents:
    The problem here is that the local M2 documentation has neither
    an HTTP header nor a <meta...> tag to set the character
    encoding, and on my system Firefox is guessing wrong.
    This change adds the tag
      <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
    inside the <head> of each HTML document, right after </title>,
    to force the UTF-8 character encoding everywhere.
Commits on Apr 8, 2016
  1. Add new package VectorFields

    bapike committed Apr 8, 2016
    It does basic computations with polynomial vector fields
    (e.g., Lie bracket, etc.).
Commits on Jan 6, 2016
  1. bump to version 0.83 of RunExternalM2

    bapike committed Jan 6, 2016
    Just update the version number.
  2. Improve an error message in RunExternalM2

    bapike committed Jan 6, 2016
    Give some reasons the error could be occurring to help out the user.
  3. Remove runExternalM2InClone function

    bapike committed Jan 6, 2016
    runExternalM2InClone forks and then does computations in the child
    M2 process.  Since M2 is multithreaded, this is at best fragile,
    and at worst terribly unsafe and unpredictable.  For instance,
    Github issue #355 shows that on newer OS X versions,
    tests for runExternalM2InClone were crashing by merely doing
    memory allocation.  Similar functionality exists in
    runExternalM2(), so just remove runExternalM2InClone().
    Also update the comment that describes how RunExternalM2 works.
  4. Remove docs for runExternalM2InClone

    bapike committed Jan 6, 2016
    Remove documentation referring to runExternalM2InClone, as it is
    about to be removed.  Also improve the documentation for
    RunExternalM2 in a few places.
  5. Remove runExternalM2InClone tests

    bapike committed Jan 6, 2016
    In prep for removing runExternalM2InClone, remove its tests.
Commits on Dec 19, 2015
  1. Change doc for fork() to be more discouraging

    bapike committed Dec 19, 2015
    Previously the documentation could be read as saying: as long as the user
    did not explicitly start new tasks, then it was OK to do stuff in the child.
    M2 always uses multiple threads, though, so it's always somewhat unsafe.
Commits on Nov 25, 2015
  1. Fix new strings tests

    bapike committed Nov 25, 2015
    These tests don't work; concatenate(-1) gives an error message, and
    pad(4,"ABC") will pad on the left.
Commits on Nov 24, 2015
  1. Quote exports in RandomIdeal

    bapike committed Nov 24, 2015
    quote exports in RandomIdeal.
  2. Quote exported symbols in SchurRingsOld.m2

    bapike committed Nov 24, 2015
    Quote exported symbols.
  3. Update syntax of SecondPackage

    bapike committed Nov 24, 2015
    Quote exported symbols, and we get complaints about 'TO "FirstPackage"'
    unless we add beginDocumentation().
  4. Update RandomSearch syntax so it loads

    bapike committed Nov 24, 2015
    Put quotes around the exported symbols, and initialize its global
  5. Split QthPower's 1 TEST into multiple TESTs

    bapike committed Nov 24, 2015
    QthPower has all tests grouped together into 1 TEST.  On a build with
    --disable-optimize --enable-debug, it took too long (811 seconds) and
    failed.  Split them into multiple TESTs that will pass.
    The last few tests seem to be coupled (or there's a typo) so I kept
    them together.
  6. Add test: substitute(1/101,ZZ/101) should error

    bapike committed Nov 23, 2015
    Check that this code, and similar code for QQ[x,y,z] and
    ZZ/101[x,y,z], give errors.
  7. Check denom!=0 when fraction is converted to ZZ/p

    bapike committed Nov 23, 2015
    The code "substitute(1/101,ZZ/101)" currently returns 0 on a
    non-debug build, and terminates in ARingZZpFlint::divide with
    "Assertion `b != 0' failed" in a debug build.  Compare that to
    "promote(1/101,ZZ/101)" which gives "error: promotion of this
    rational number to the ring R not possible" on both types of
    Change the behavior of ARingZZpFlint::set_from_mpq to error
    if the denominator was zero, before it tries to do division in
    ZZ/p.  I'm not positive I'm handling the error correctly.
    This was noticed because packages/Macaulay2Doc/test/subst5.m2
    was failing on debug builds.
  8. Re-remove test that consumes too much memory

    bapike committed Nov 23, 2015
    This test consumes a lot of memory.  For details, see commit
    ab33ead from Aug 2014, which
    originally removed this test.  It was re-added in commit
    2af61bd or
    189d50c, probably by mistake,
    in an update to MultiplierIdeals.
  9. Relax this speed test so debug 'make check' works.

    bapike committed Nov 22, 2015
    In a debug build on my system, this test took 0.37*standardSecond,
    whereas on the released 1.8.2 it takes 0.227*standardSecond.
    Relax the standard a little.
  10. Ignore certain Debug messages that can make tests from the Computatio…

    bapike committed Nov 22, 2015
    …nsBook fail.
    Only the "changing factory characteristic" message was observed, from a test in geometry/.
    The code that produces the message, in M2/x-factor.cpp, also could conceivably produce
    the other messages, too.
  11. Add tests for concatenate(numbers), pad

    bapike committed Nov 24, 2015
    Test that adjacent commit stays fixed.
Commits on Nov 20, 2015
  1. Fix a miscalculation of the expected string length

    bapike committed Nov 20, 2015
    In a debug build, concatenate("ABC",-1) will crash M2 because it miscalculates
    the expected length of the final string (concatenate("ABC",2) gives "ABC  ").
    Negative numbers like this arise by doing something like pad("ABC",2);
    though this change makes it just return the whole string, reasonable behavior could also
    be to truncate the string.
    This was causing a crash in the "minprimes" example inside the MinimalPrimes package.
  2. Backport a fix from MPIR issue #12 to try to fix M2 issue #35.

    bapike committed Nov 19, 2015
    This should make sqrt 2p3000000 work, hopefully.
Commits on Nov 19, 2015
Commits on Sep 4, 2015
  1. On OS X, get("!/usr/bin/time --version") gives an error because time …

    bapike committed Sep 4, 2015
    …has a nonzero exit code. Use run() instead everywhere.
Commits on Sep 3, 2015
  1. Bump to version 0.82

    bapike committed Sep 3, 2015
  2. Change default PreRunScript setting.

    bapike committed Sep 3, 2015
    A shell on OS X doesn't know the -n param to echo, so "echo -n" results in showing a "-n".
Commits on Aug 24, 2015
Commits on Jul 15, 2015
  1. v0.81 of RunExternalM2:

    bapike committed Jul 15, 2015
      o remove setExternalM2Child() and use Configuration option instead
      o change default M2Location to null so that documentation looks OK
      o use toAbsolutePath when finding default M2Location
Commits on Jun 10, 2015
  1. Make documentation for temporaryFileName() more precise about file re…

    bapike committed Jun 10, 2015
    …moval behavior if fork is used.
You can’t perform that action at this time.